Automated Classification Schemes for Optical Tomographic Arthritis Scans

We have recently developed a sagittal laser optical tomographic (SLOT) imaging system for the diagnosis and monitoring of inflammatory processes in proximal interphalangeal (PIP) joints of patients with rheumatoid arthritis (RA). While cross sectional images of distribution of optical properties can now be generated easily, clinical interpretation of these images remains a challenge. In this paper, we apply and analyse two machine learning methods for optimal identification and severity classification of RA in a data set of 78 joints. The methods surveyed include fisher face with support vector machines (SVMs), and transformed mixtures of Gausians (TMG). It appears that TMG methods outperform the approach using fisher face with SVMs; however, the results need to be further validated in studies involving larger patient populations